Abstract
Background. There are large geographic differences in breast cancer risk but whether survival differs between low- and high-risk groups is less well-established. As the survival of cancer depends on the level of healthcare and awareness of disease risks, subtle differences in cancer biology cannot be revealed in international comparisons. Instead, comparison of diverse immigrant groups in a country of uniformly accessible healthcare system should enable conclusions to be made about ethnic determinants of cancer risk and survival. Material and methods. The Swedish Family-Cancer Database was used to calculate standardized incidence (SIRs) and hazard ratios (HRs) of death from female breast cancer in 12 505 and 137 547 patients diagnosed with breast cancer among immigrants and Swedes, respectively. The ratios were adjusted for age, period, region, parity, and age at first childbirth. Ordinal logistic regression analysis was used to estimate odds ratios (ORs) for the clinical TNM classes. The analyses were stratified by menopausal status and histology. Results. Turks, Southeast Asians, and Chileans had the lowest breast cancer risk (SIR = 0.44; 95% CI 0.37–0.51) and Iraqis the highest risk (1.19; 1.05–1.35), mainly due to premenopausal cancer (1.51; 1.27–1.78). The HRs for all breast cancers were between 0.98 (0.81–1.18) (low-risk Europeans) and 1.24 (0.94–1.63) (lowest-risk non-Europeans), but were not significant. No differences in survival of ductal carcinoma between immigrants and Swedes were found, while low-risk non-Europeans had a HR of 2.88 (1.37–6.08) for lobular carcinoma. Low-risk non-Europeans were diagnosed in a higher T-class (OR = 1.87; 1.21–2.87) than Swedes. Conclusion. We did not find any evidence that ethnic differences in breast cancer risk substantially affect the survival. The observed poor survival of some low-risk immigrants in lobular carcinoma may be related to treatment. The tendency of low-risk immigrants to present with higher T-class compared to Swedes may depend on their lower participation in the mammography screening program.
The global incidence in breast cancer varies extensively from the lowest rate of 19.3 per 100 000 in Eastern Africa to the highest rate of 89.9 per 100 000 in Western Europe [Citation1,Citation2]. The data on survival differences are more limited, covering ethnic groups in the USA or national estimates between European countries and Australia. A study on a five-year relative survival in the 1990s, based on cancer registries from 31 countries, showed the best survival rates for Cuba, USA, Canada, Sweden and Japan (well over 80%) and the worst ones for Algeria (39%), Slovakia and Brazil (both 58%), Estonia (61%), and Czech Republic (63%) [Citation3]. In some countries, such as Algeria and Brazil, cancer registration only covers a small proportion of the population which may weaken the generalizability of the results. However, the Algerian results were approximately similar to the ad hoc survival study from Harare, Zimbabwe, collected around the mid-1990s [Citation4]. Even though direct survival data are sparse from the developing countries, these can be estimated through incidence and mortality data, taking advantage of the fact that mortality to incidence ratio is approximately equal to one minus long-term relative survival [Citation2]. For example, relating to the above incidence figures showing a 4.7-fold difference between Western Europe and Eastern Africa, the mortality difference was only 1.5-fold, implying considerable survival advantage for Western Europeans [Citation2]. The mortality figures for breast cancer through most of Africa are higher than those for Western Europe. They are even slightly higher in most of the developing world with low incidence rates, indicating much worse survival in those countries compared to Western Europe [Citation2].
The differences in breast cancer survival are thought to be largely related to access to diagnostic and treatment services and to overall health awareness and concerns of the populations [Citation3]. However, there appear to be biological differences in ethnic presentation of breast cancer and these may also contribute to the survival differentials. While breast cancer is mainly a postmenopausal disease in Western countries with ductal carcinoma accounting for 2/3 and lobular carcinoma 1/10 of all cases, premenopausal breast cancer with overwhelming ductal histology is more common in developing countries [Citation5]. These differences appear to be maintained in immigrants migrating from low- to high-risk countries [Citation6,Citation7]. The data on an early age of onset and prevalent ductal histology were also demonstrated among immigrant women in Sweden [Citation7].
Whether biological ethnic differences contribute to survival differences cannot be concluded from international comparisons because of the differences in healthcare systems. Also in countries with a multilevel, socially segregated healthcare system, such as the USA, reasons for the ethnic differences are difficult to unravel. The immigrant population in Sweden may allow a more definitive study because the healthcare system is entirely public at a nominal cost and the country was the first to offer a national mammographic screening program [Citation8]. Furthermore, the immigrants in Sweden make up some 15% of the population and they originate from practically all countries around the world. In the present study, we focus on survival in female breast cancer among immigrants and want to resolve the question of whether different levels of breast cancer risk also imply differences in survival. We have previously reported breast cancer risk in immigrants but the data on survival is entire novel [Citation7]. Standardize incidence ratios (SIRs) show the incidence differences between immigrants and native Swedes in breast cancer adjusted for a number of variables. Survival differences are given by hazard ratios (HRs). The HRs measure the timing of breast cancer deaths in immigrants compared to Swedes. In addition to total breast cancer, pre- and postmenopausal types and ductal and lobular histologies, the most common histological subtypes, are separately analyzed [Citation9]. In order to interpret the results, data are given on the clinical tumor node metastasis (TNM) classes.
Methods
Study design and patients
The Family Cancer Database (FCD) contains individuals born in Sweden since 1932 with their biological parents. It also contains first-generation immigrants that were defined as those born outside Sweden without identified parents in the dataset. This dataset includes population-based data from multigenerational registries, national censuses (1960, 1970, 1980, and 1990), the Swedish Cancer Registry, and death notifications [Citation10]. Native Swedes were defined as those who, along with their parents, were born in Sweden. Based on our previous study and to catch the maximum number of deaths, we classified the countries according to geographical setting and breast cancer risk into high-risk immigrants (Iraqis), low-risk Europeans (Balts and Greeks), low-risk non-Europeans (0.60 ≤ SIR < 0.80) (Indians, East Asians, Latin Americans, and other Africans), the lowest-risk non-Europeans (SIR < 0.60) (Turks, Southeast Asians, and Chileans), and other immigrants (Finns, Danes, Norwegians, Germans, Benelux, Britons, Poles, East and South Europeans, Russians, former Yugoslavians, Iranians, Asian Arabs, North Americans, North Africans, Iceland, Andorra, Cyprus, Israel, Malta, Monaco, Portugal, San Marino, State of the Vatican City, Albania, Macedonia, Moldova, Slovenia, Afghanistan, Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Ukraine, Uzbekistan, and Belarus) [Citation7]. The last group was the largest one with immigrants from many different ethnicities and with a similar risk of breast cancer as in Swedes.
Following the start of cancer registration in 1958, data on cancers in the FCD had a four-digit diagnostic code according to the seventh revision of the International Classification of Disease (ICD). The codes for breast cancer were 170.1 and 170.2. Since 1993, ICD-O-2/ICD with histopathological data according to the Systematized Nomenclature of Medicine (SNOMED) has been used. This coding system gives a detailed tumor histology-topology. We analyzed ductal (SNOMED = 8500 and 8501) and lobular (8520) carcinomas. Due to the small number of cases among immigrants, non-specific adenocarcinomas (SNOMED = 8140; N = 275), tubular (8211; 132), and mucinous (8480; 0) carcinomas were not included in our histology-specific analysis. All breast cancer cases (100%) registered in the Cancer Registry were histologically verified [Citation11].
Statistical analysis
The SIRs were calculated as the ratio of observed to expected number of cases, using native Swedes as the reference population [Citation12]. The expected number of cases was calculated in five-year age groups and in 10-year intervals for the years between 1958 and 2008. The start of the follow-up was defined as birth year, the date of immigration, or 1 January 1958, whichever was the latest. The end of the follow-up was defined as the date of cancer diagnosis or emigration, end of the last year of presence for the census, or the closing date of our study, 31 December 2008, whichever was the earliest. The SIRs were adjusted for age (five-year age groups), period (10-year bands from 1958 to 2008), region (Northern, Southern, and other Swedish regions), parity, and age at first childbirth (< 20 years, 20–29 years, and ≥ 30 years) [Citation7]. The confidence interval (95% CI) for the ratios was calculated by making a Poisson distribution. We also calculated 99% CI for the ratios to reduce chance findings [Citation13]. The associations were reported according to 0.05 and 0.01 significance levels.
To compare site-specific cancer survival between immigrants and Swedes, the HRs were calculated using a Cox model (PROC PHREG; SAS software version 9.2; SAS Institute Inc., Cary, NC, USA). The underlying time scale was the time since diagnosis in the analyses of fatal site-specific cancer. Individuals entered the risk period at the date of breast cancer diagnosis. The event was death due to breast cancer. The censoring event was the end of the follow-up, which was defined as the closing date of study, emigration, and death due to other causes than breast cancer, whichever was the earliest. Age, period, region, parity, and age at first childbirth were taken into account as covariates. The proportional hazard assumption for the covariates was tested by Schoenfeld residuals and by plotting the log of the negative log of the survival function versus the log of time.
Tumor size and local growth (T), regional lymph node involvement (N) and the presence of metastasis (M) according to the TNM classification system published by the American Joint Committee on Cancer have been available in the FCD since 2002 [Citation14,Citation15]. An ordinal logistic regression was used to calculate odds ratios (ORs) (PROC LOGISTIC; SAS software version 9.2; SAS Institute Inc., Cary, NC, USA) for breast cancer according to the TNM classification, using T1, N0 and M0 as the references for ‘T2–T4’, ‘N1–N3’, and ‘M1’, respectively. The ORs were adjusted for age at diagnosis (< 50 years, 50–59 years, 60–69 years, and ≥ 70 years), region, parity, and age at first childbirth. The 95% CI was calculated. A p-value of less than 0.05 was considered statistically significant. Due to the small numbers of M1 cases (Swedes: N = 636; immigrants: N = 77), this analysis was only presented for T and N cases. Because of the small number of cases with in situ cancer or unknown TNM classes, we also excluded Tis, TX, NX, and MX from our analysis. To assess the effect of the median age at menopause, the analyses were additionally stratified by age (≤ 50 and > 50 years) [Citation16,Citation17].
Results
Our dataset included 137 547 patients and 36 472 deaths from breast cancer in Swedes, and 12 505 patients and 2765 deaths in immigrants (). The SIR for all breast cancer was 0.54 among the lowest-risk non-European immigrants (Turks, Southeast Asians, and Chileans), compared to Swedes with the reference SIR of 1.00. The SIR was increased to 1.19 for the high-risk immigrants (Iraqis). The HRs were similar and not statistically significant: 1.24 and 1.22, respectively. For postmenopausal breast cancer, the SIR was 0.44 and the HR was 1.37 for the lowest-risk non-Europeans. Among high-risk immigrants, the HR was non-significantly increased to 1.92. For premenopausal breast cancer, the SIR was 1.51 for high-risk immigrants, and the HR was significantly increased to 1.11 in the group of other immigrants. The difference in HRs between the best and poorest survivor groups for all breast cancer was 1.27-fold.
Table I. Standardized incidence and hazard ratios (SIRs and HRs) for breast cancer among immigrant women in Sweden by menopausal status from 1958 to 2008.
High-risk immigrants showed an SIR of 1.24 for ductal carcinoma but the HRs were not significant for any immigrant group (). The lowest-risk non-Europeans had also the lowest risk of lobular carcinoma (0.29) and even most other immigrant groups had low risks. By contrast, low-risk non-Europeans (Indians, East Asians, Latin Americans, and other Africans) had 2.88-fold increase in HR, while the increased HR for high-risk immigrants (2.76) was not significant. Further analysis of the histological types in pre- and postmenopausal cancer showed the only significant increase in HR for lobular histology in premenopausal cancer among low-risk non-Europeans (N = 5, HR = 3.03, 95% CI 1.11–8.28); among these, two were from the Indian subcontinent (HR = 5.22, 1.23–21.63) and two from Latin America (HR = 8.54, 2.05–35.53) (data not shown).
Table II. Standardized incidence and hazard ratios (SIRs and HRs) for breast cancer among immigrant women in Sweden by histology from 1993 to 2008.
Data for T and N classes are shown in . We found for all breast cancer that low-risk non-Europeans were diagnosed in a higher T-class (OR = 1.87; T2–T4 compared to T1) than Swedes and the OR was particularly high (2.62) for premenopausal patients. Among the lowest-risk non-Europeans a higher T-class was noted for all breast cancer (1.79) and for postmenopausal cancer (2.55). Further analysis by histology showed that both of the increases were significant in ductal carcinomas: Low-risk non-Europeans (N = 45, OR = 2.68, 1.35–5.35) and lowest-risk non-Europeans (N = 26, OR = 2.15, 1.47–6.59). We found no difference in the risk of lymph node involvement between immigrants and Swedes (data not shown).
Table III. Odd ratios (ORs) for breast cancer according to the T and N classification by birth region and menopausal status from 2002 to 2008.
Discussion
The novel result of the present study was that there were no large differences in breast cancer survival between the ethnic groups in spite of the large differences in breast cancer incidence. In fact, the observed 1.27-fold difference in HRs between the best and poorest survivor groups was less than 1.47-fold observed between Swedish women by educational level [Citation18]. These data imply that entirely different population-level risk factors regulate risk and survival, also illustrated by the fact that in the past 40 years age-standardized incidence in breast cancer increased 1.8-fold in Sweden while five-year relative survival improved from 60% to 85% [Citation19]. The histological differences in breast cancer presentation between immigrants and Swedes are too small to influence survival, once age at diagnosis is considered [Citation7,Citation20].
An exception to the uniform survival was the poor survival for lobular breast cancer among low-risk non-Europeans (2.88), mainly premenopausal patients (3.03). Lobular cancers are not easily detected in mammography and it is not likely that non-participation in mammography screening explains the results because in most immigrant dense areas of Sweden, such as large cities, mammography invitations to organized screening start at the age of 50 years, i.e., at postmenopausal age [Citation21–23]. Lobular carcinoma is frequently hormone receptor positive and it is responsive to adjuvant endocrine therapy [Citation24]. Thus one possible reason for the poor survival could be lower prescription of or lack of compliance with adjuvant therapy. However, why Indians and Latin Americans with the highest HRs in premenopausal lobular cancer would be particularly affected cannot be explained. The data on higher T-class among low-risk non-European immigrants may be at least partially related to participation in mammographic screening because the largest increases were found in postmenopausal ductal cancers.
A curiosity in the data was the high risk of breast cancer in Iraqi women even though Globocan predicts a relatively low risk in the Near East countries [Citation1,Citation2]. We have, however, found in earlier studies that the breast cancer incidence in immigrants from other Asian Arab countries and Iran is close to that of Swedes and there is a tendency for the incidence to approach the Swedish rate with longer duration of residence [Citation3]. Already the early immigrant study on Italian women to Australia on breast cancer found an effect of residential time: the mortality in Italian women increased from 50% to over 100% compared to Australians when 17 years had passed after immigration [Citation25]. Consequently, the high incidence in Iraqi women appears not to be a coincidence and the apparent contradiction with the Globocan data may have several explanations, such as the incidence rates may have increased more for Iraqi women after immigration than for other women from developing countries, Globocan underestimates the incidence in the Near East, or immigrants are a selective group of the original population [Citation3].
In summary, our findings indicate that the ethnic differences in breast cancer risk had no major influence on survival. The observed poor survival of some low-risk immigrant groups in lobular carcinoma may be related to treatment. The tendency of low-risk immigrants to present with higher T-class compared to Swedes may depend on their lower participation in the mammography screening program. The high risk of breast cancer in Iraqi women shows that immigrants from assumed low-risk areas may in fact not be or remain at low risk.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
This study was supported by the Deutsche Krebshilfe, the Swedish Council for Working Life and Social Research and EU FP7/2007-2013 grant 260715.
References
- Ferlay JS, Bray F, Forman D, Mathers C, Parkin DM. GLOBOCAN 2008, Cancer incidence and mortality worldwide: IARC Cancer Base No. 10. [database on the Internet]. International Agency for Research on Cancer. 2010 [cited 2010]. Available from: http://globocan.iarc.fr. 2010.
- Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893–917.
- Coleman MP, Quaresma M, Berrino F, Lutz JM, De Angelis R, Capocaccia R, et al. Cancer survival in five continents: A worldwide population-based study (CONCORD). Lancet Oncol 2008;9:730–56.
- Gondos A, Chokunonga E, Brenner H, Parkin DM, Sankila R, Borok MZ, et al. Cancer survival in a southern African urban population. Int J Cancer 2004;112:860–4.
- Tavassoli F, Devilee P. Tumours of the breast and female genital organs. Lyon: IARC Press; 2003.
- Mousavi SM, Fallah M, Sundquist K, Hemminki K. Age- and time-dependent changes in cancer incidence among immigrants to Sweden: Colorectal, lung, breast and prostate cancers. Int J Cancer 2012;131:E122–8.
- Hemminki K, Mousavi SM, Sundquist J, Brandt A. Does the breast cancer age at diagnosis differ by ethnicity? A study on immigrants to Sweden. Oncologist 2011;16:146–54.
- Sultman RB, Busse R, Figueras J. Social health insurance system in Western Europe: World Health Organisation 2004. Maidenhead: Open University Press; 2004.
- Boyle P, Levin B. World Cancer Report 2008. Lyon: IARC; 2008.
- Hemminki K, Ji J, Brandt A, Mousavi SM, Sundquist J. The Swedish Family-Cancer Database 2009: Prospects for histology-specific and immigrant studies. Int J Cancer 2009; 126:2259–67.
- The National Board of Health and Welfare. Cancer Incidence in Sweden, 2010. Socialstyrelsen: Stockholm; 2010.
- Mousavi SM, Sundquist K, Hemminki K. Morbidity and mortality in gynecological cancers among first- and second-generation immigrants in Sweden. Int J Cancer 2012;131: 497–504.
- Stigler S. Fisher and the 5% level. Chance 2008;21:12.
- Greene FL, Page DL, Fleming ID, Fritz A, Balch CM, Haller DG, et al. AJCC (American Joint Committee on Cancer) cancer staging manual, 6th ed. New York: Springer-Verlag; 2002.
- Cowherd SM. Tumor staging and grading: A primer. Methods Mol Biol 2012;823:1–18.
- Rodstrom K, Bengtsson C, Milsom I, Lissner L, Sundh V, Bjourkelund C. Evidence for a secular trend in menopausal age: A population study of women in Gothenburg. Menopause 2003;10:538–43.
- Hidayet NM, Sharaf SA, Aref SR, Tawfik TA, Moubarak II. Correlates of age at natural menopause: A community-based study in Alexandria. East Mediterr Health J 1999;5:307–19.
- Hussain SK, Altieri A, Sundquist J, Hemminki K. Influence of education level on breast cancer risk and survival in Sweden between 1990 and 2004. Int J Cancer 2008;122:165–9.
- Engholm G, Ferlay J, Christensen N, Bray F, Gjerstorff ML, Klint Å, et al. NORDCAN: Cancer incidence, mortality, prevalence and prediction in the Nordic countries, Version 3.6. Association of the Nordic Cancer Registries. Danish Cancer Society. 2010 [cited 2012 Nov]. Available from: http://www-dep.iarc.fr/NORDCAN/english/frame.asp
- Mousavi SM, Zheng T, Dastgiri S, Miller AB. Age distribution of breast cancer in the middle East, implications for screening. Breast J 2009;15:677–9.
- Lagerlund M, Maxwell AE, Bastani R, Thurfjell E, Ekbom A, Lambe M. Sociodemographic predictors of non-attendance at invitational mammography screening – a population-based register study (Sweden). Cancer Causes Control 2002;13: 73–82.
- Olsson S, Andersson I, Karlberg I, Bjurstam N, Frodis E, Hakansson S. Implementation of service screening with mammography in Sweden: From pilot study to nationwide programme. J Med Screen 2000;7:14–8.
- Rakha EA, El-Sayed ME, Powe DG, Green AR, Habashy H, Grainge MJ, et al. Invasive lobular carcinoma of the breast: Response to hormonal therapy and outcomes. Eur J Cancer 2008;44:73–83.
- Arpino G, Bardou VJ, Clark GM, Elledge RM. Infiltrating lobular carcinoma of the breast: Tumor characteristics and clinical outcome. Breast Cancer Res 2004;6:R149–56.
- McMichael AJ, Giles GG. Cancer in migrants to Australia: Extending the descriptive epidemiological data. Cancer Res 1988;48:751–6.